-
Bug
-
Resolution: Unresolved
-
Undefined
-
None
-
4.18.z
-
None
-
None
-
False
-
-
None
-
Moderate
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Description of problem:
Controller-manager pod has some routines by default described below:
controllers:
- openshift.io/build
- openshift.io/build-config-change
- openshift.io/builder-rolebindings
- openshift.io/builder-serviceaccount
- -openshift.io/default-rolebindings
- openshift.io/deployer
- openshift.io/deployer-rolebindings
- openshift.io/deployer-serviceaccount
- openshift.io/deploymentconfig
- openshift.io/image-import
- openshift.io/image-puller-rolebindings
- openshift.io/image-signature-import
- openshift.io/image-trigger
- openshift.io/ingress-ip
- openshift.io/ingress-to-route
- openshift.io/origin-namespace
- openshift.io/serviceaccount
- openshift.io/serviceaccount-pull-secrets
- openshift.io/templateinstance
- openshift.io/templateinstancefinalizer
- openshift.io/unidling
The "openshift.io/image-signature-import" one is making every 1 hour many requests against quay.io to check for signatures that is spamming environments that have a proxy or disconnected without any (from my understanding) valid reason.
This pod does not even have credentials (access to the global pull-secret) so this routine does not make sense to exist.
Also im not sure what is the actual reason behind this implementation TBH.
I was able to disable this controller by setting the controller-manager operator to unmanaged, scaling the deployment down, editing the "config" configmap by adding a "-" in front of the controller in the json and restarting all the controller-manager pods manually. But this is unsupported procedure
Version-Release number of selected component (if applicable):
How reproducible:
Simply enable the traceALL log verbosity on the openshift-controller-manager pods and all the requests will appear.
Steps to Reproduce:
1. Enable TraceAll verbosity on the controller-manager pods
2. Find the controller-manager pod leader
3. Check the logs for "quay.io"
Actual results:
The controller-manager pod is checking signatures from public quay.io
Expected results:
The controller-mananer should not ?? check anything from public quay.io
Additional info:
Please check if this feature/functionality is important to at least make an RFE to maybe allow disabling this in a supported way from the Operator CR.